Thoughts on access control, AI governance, and the security problems worth solving.
Lutril generates one installer per platform with your token in it; push it with Intune, Jamf, Kandji or a terminal and the fleet enrols itself. The settings that matter, what the script does, how to confirm the rollout landed, and how to remove it. No package to sign.
How to give AI agents temporary access that ends on its own: runtime decisions at the tool call, agent access policies, one-call grants, automatic revocation.
Just-in-time access grants permission for a fixed window and revokes it on its own. Why standing access keeps showing up in the 2026 breach reports and how to replace it with time-boxed, approved grants.
OAuth discovery surfaces every app an employee authorized with Sign in with Google or Microsoft. It cannot see the tools people sign up for with a work email and a password. Our new browser extension closes that gap, privacy first.
MCP governance is how you control, authorize, and audit what AI agents can do in your SaaS tools through the Model Context Protocol. What it is, why API keys aren't enough, the MCP gateway pattern, and a checklist to govern AI agent access.
Most employees already use AI tools your security team never approved. Shadow AI is shadow IT that reads your data and acts on your systems. Here is why it spreads, what it costs, and how to govern it instead of banning it.
Your security stack controls who can access your SaaS tools. It has no idea your AI agents even exist. The problem, the gaps, and what governance for agents actually looks like.
AI agents need identity, policy enforcement, audit logs, and a kill switch before they touch your SaaS stack. That infrastructure has a name. Here is what it is and how it works.
How to audit AI agents after employee offboarding: find every agent the leaver owned, then reassign, suspend or revoke each one with a logged decision. Why disabling the Okta account leaves agents running.
Need an AI agent governance layer that works with your existing Okta or Azure AD? What each IdP covers for agents, the five gaps both leave open, and how a governance layer plugs in on top.
Step-by-step: register a Claude agent in Lutril, connect Slack as an MCP app, bind the integration with read/write scope, and wire it into Claude. Every tool call logged, kill switch included.
Disabling an Okta account is not offboarding. 41% of employees keep access to company systems after leaving. Here is what complete access revocation actually requires.
The average company uses 3x more SaaS tools than IT knows about. Shadow IT is not a policy problem. It is a structural one. Here is why it happens, what risk it creates, and how to manage it.
Most access reviews are a formality. Managers approve without context, shadow IT goes unreviewed, and evidence is thin. Here is what SOC 2 and ISO 27001 actually require.